This invention generally relates to laser diodes and further to the art of optical telecommunication that uses a laser diode. Especially this invention is related to a so-called surface-emission laser diode that emits a laser beam in a generally vertical direction to a substrate surface. Also, the present invention is related to an optical transmission/reception system and optical-fiber telecommunication system that uses such a surface-emission laser diode. Further, the present invention relates to a semiconductor distributed Bragg reflector and also a surface-emission laser diode and further a surface-emission laser array. Further, the present invention relates to a surface-emission laser module, an optical interconnection system and an optical telecommunication system.
A surface-emission laser diode is a laser diode that emits a laser beam in a generally vertical direction from a surface of a substrate. By using surface-emission laser diodes, two-dimensional array integration of laser diode is achieved easily. Further, the laser diode has an advantageous feature of relatively narrow divergent angle of the output optical beam (about 10 degrees), which is particularly suitable for coupling with optical fibers. Furthermore, inspection of the laser diode device is made easily in a surface-emission laser diode.
Thus, surface-emission laser diodes are suited to construct an optical transmission module (optical interconnection apparatus) of parallel-transmission type, and research and development are conducted prosperously. The immediate application of the optical interconnection apparatus would be the parallel connection between computers or circuit boards in a computer, including short-range optical-fiber telecommunication. In future, application to a large-scale computer network and trunk line system of long-range, large-capacity telecommunication is expected.
Generally, a surface-emission laser diode includes an active layer of a group III-V semiconductor material such as GaAs or GaInAs, and an optical resonator is formed by disposing an upper semiconductor Bragg reflector and a lower semiconductor Bragg reflector arranged respectively above and below the active layer.
In such a construction, the length of the optical resonator is remarkably short as compared with the case of an edge-emission laser diode. Therefore, it is necessary to increase the reflectance of the reflector to a high value (99% or more) for facilitating laser oscillation. Because of this, it is practiced to use a semiconductor Bragg reflector in a surface-emission laser diode as a reflector, wherein a semiconductor Bragg reflector is formed of an alternate and repetitive stacking of a high-refractive index material such as GaAs and a low refractive index material such as GaAs with a period of ¼ wavelength.
However, in the conventional semiconductor Bragg reflector that has the structure mentioned above, there arises a spike structure in the energy band as a result of band discontinuity at the hetero interface, at which the materials of different bandgaps are jointed, and the spike structure thus formed tends to function as a barrier against carriers. Thereby, there arises a problem in that the semiconductor multilayer part increases the resistance of the laser diode. Because of this, conventional surface-emission laser diodes constructed on a GaAs substrate have suffered from the problem of comparatively high operating voltage of about 2.5 volts or more. Because of this, it has been difficult to use the surface-emission laser diode with a CMOS driver integrated circuit, which produces a laser driving voltage of 2 volts at best. The itemization of this operating voltage of 2.5 volts is: 1.5V for the diode part; and 1V for the device resistance. In order to reduce the operational voltage below 2 V, it is necessary to reduce the device resistance by one-half, while it is extremely difficult to meet for this requirement at the present stage of technology.
In the case of a laser diode of long-wavelength band for use in optical telecommunication, such as the laser diode of 1.3 μm band or 1.55 μm, a low voltage operation is expected in view of the fact that only a voltage of 1 volt or less is applied to the diode part of the laser diode. Unfortunately, the desired low voltage operation is not materialized in such a long wavelength laser diode. In conventional long-wavelength laser diode, InP is used for the substrate and InGaAsP is used for the active layer. In such a system, the lattice constant of InP constituting the substrate is large, and it is difficult to achieve a large refractive-index difference in the reflector when a material that achieves lattice matching with the InP substrate is used for the reflector. Consequently, it has been necessary to stack 40 or more pairs in the reflector for realizing sufficient reflectance. In such a construction, however, the resistance of the reflector increases again as a result of increased stacking number of the reflector. Thus, it has been difficult to drive the laser diode driver by a CMOS integrated circuit.
In a surface-emission laser diode formed on an InP substrate, there is another problem of change of laser characteristic caused by the temperature. Because of this, it has been necessary to add an apparatus for stabilizing the temperature in the laser diode constructed on such an InP substrate. However, the use of such a temperature regulator is difficult in the apparatus for home use, which is subjected to a severe demand of cost reduction. Because of these problems of increased number of stacking in the reflector and the poor temperature characteristics, practical long-wavelength surface-emission laser diode has not yet commercialized.
In order to deal with the foregoing problems, there is a proposal to construct a surface-emission laser diode on a GaAs substrate by using an AlInP layer, which achieves a lattice matching with the GaAs substrate, in at least one of the upper and lower semiconductor Bragg reflectors as the low refractive index layer, and further by using a GaInNAs layer in at least one of the upper and lower semiconductor Bragg reflectors, as disclosed in Japanese Laid-Open Patent Application 9-237942, such that a large refractive index difference is realized in the reflector and the number of stacking therein is reduced while maintaining high reflectance.
In the foregoing conventional art, the bandgap of the active layer is reduced by 1.4 eV by using GaInNAs, in which N is introduced into the III-group V semiconductor material system of GaInAs. As a result, the laser diode can produce an optical beam with a wavelength longer than 0.85 μm. In the aforementioned prior art, it should be noted that the material system of GaInNAs can achieve a lattice matching with the GaAs substrate. Further, the prior art describes the semiconductor layer of GaInNAs can be a promising material for the long-wavelength surface-emission laser diode operable in the 1.3 micron band and 1.55 micron band.
In spite of such a description in the prior art with regard to the possibility of surface emission laser diode operable in the wavelength band longer than 0.85 μm, there has no such a laser diode actually materialized. The present situation would be something like that the theoretical construction is already established but the actual construction for materializing the laser diode is not discovered yet.
In one example, there is a laser diode that uses a semiconductor Bragg reflector formed by stacking high-refractive index material layers of GaAs and low refractive index material layers of AlAs alternately as noted above with the periods of ¼ wavelength. However, the laser diode structure thus formed does not provide optical emission at all, or operates but only with low power, indicating that the efficiency of optical emission is extremely small.
Similarly, there is a laser diode disclosed in the Japanese Laid Open Patent Application 9-237942 in which an AlInP layer is used for the low refractive index layer of the semiconductor Bragg reflector. In this case, too, the luminous efficacy of the laser diode is far from the level of practical use.
The reason of this unsatisfactory result is attributed to the chemical activity of the material including Al. More specifically, it is thought that the use of a material containing Al easily invites formation of crystal defects originating from Al. Thus, there have been proposals, as in the Japanese Laid-Open Patent Application 8-340146 and Japanese Laid-Open Patent Application 7-307525, to construct the semiconductor Bragg reflector with materials free from Al such as GaInNP and GaAs. However, the material system of GaInNP and GaAs can provide a refractive-index difference of about half as compared with the material system of AlAs and GaAs. Thus, the stacking number in the reflector has to be increased, and the object of reducing the resistance of the surface-emission laser diode is not attained.
Thus, at present, the surface-emission laser diode operable at the long-wavelength of 1.1-1.7 μm does not exist, and because of this, it is not possible to construct a computer network or optical-fiber telecommunication system that uses such a laser diode.
As explained before, in a conventional surface-emission laser diode, it was also not possible to use a CMOS circuit for the laser diode driver, and it has been necessary to use an expensive special driver circuit. On the other hand, if a mass-produced CMOS driver integrated circuit could be used, the cost of the optical telecommunication system that uses such a surface-emission laser diode would be reduced significantly.
Furthermore the use of a CMOS circuit can reduce the power supply voltage of the driver integrated circuit as well from 5V to 3.3V. With this, it is possible to reduce the power consumption of the system to about one-half, and a very large effect of electric power saving is obtained.
As noted before, there is a widespread expectation of optical-fiber telecommunication in relation to computer networks, and the like. Especially, there is a need of realizing a low cost system in order that the public accepts such an optical telecommunication system. Unfortunately, the surface-emission laser diode that can be used for this purpose and can be used with a low-cost CMOS driver integrated circuit, and oscillates at the long-wavelength band of 1.1-1.7 μm does not exist. Hence, the telecommunication system that uses such a surface emission laser diode does not exist.
Meanwhile, in the abovementioned semiconductor Bragg reflector, in which semiconductor layers of different bandgaps are grown alternately, there arises the problem of spike formation in the band structure thereof at the hetero interface as a result of the band discontinuity. When such a spike structure is formed, the spike structure functions as a barrier with regard to the carriers. Thus, there arises a problem in that the electric resistance becomes very high in the semiconductor multilayer part of the surface-emission laser diode. This effect also contributes to the large drive voltage of 2.5 V for the surface-emission laser diode constructed on a GaAs substrate. As noted previously, it has been difficult to drive a laser diode having such a large driving voltage by the driver integrated circuit formed of a CMOS circuit (driving voltage is below 2 volts).
As noted previously, the itemization of this operating voltage of 2.5 volts is: 1.5V for the diode part; and 1V for the device resistance, and it is necessary to reduce the device resistance by one-half in order to drive the laser diode with a drive voltage below 2 volts. However, this is a very difficult subject.
Recently, the optical systems are used also for peripheral transmission/reception systems, and there is a widespread expectation about the computer networks using the optical-fiber telecommunication technology including such a peripheral transmission/reception system. Especially, there is a keen interest about a low cost optical system required for spreading of the optical fiber technology to the general public. However, the surface-emission laser diode that can be used for this purpose and can be used with a low-cost CMOS driver integrated circuit, and oscillates at the long-wavelength band of 1.1-1.7 μm does not exist yet. Hence, the telecommunication system that uses such a surface emission laser diode does not exist at the moment.
In such an optical-fiber telecommunication system that uses the long-wavelength surface-emission laser diode operating at the wavelength band of 1.1-1.7 μm, the photodetection device constructed on a Si substrate cannot be used, as such a photodetection device cannot detect the wavelength of 1.1-1.7 μm. In such a system, it is necessary to use a photodetection device that has a sensitivity to the wavelength of 1.1-1.7 μm. However, the photodetection device that has sensitivity to the desired wavelength band of 1.1-1.7 μm is expensive as compared with the low cost Si photodetection device. Thus, simple replacement of a conventional Si photodetection device with the photodetection device having the sensitivity to the wavelength of 1.1-1.7 μm causes an increase of cost of the whole optical-fiber telecommunication system. Thus, in order to realize an optical telecommunication system that uses the long-wavelength surface-emission laser diode of 1.1-1.7 μm band, an approach other than replacing the conventional Si photodetection device with an expensive photodetection device is needed.
Furthermore, a GaInNAs active layer having a high strain is used in the long-wavelength surface-emission laser diode, as will be explained below. In such a laser diode, deterioration of device characteristic may be caused as a result of the thermal stress caused by the difference of linear thermal expansion coefficient with regard to the mounting substrate.
Meanwhile, in the optical-fiber telecommunication system that uses a surface-emission laser diode, it is possible to arrange a number of laser diode elements, each formed of a surface-emission laser diode, with high integration density. Thus, the distance between the optical fibers can be reduced as compared with the case in which a conventional edge-emission laser diode is used for the laser diode array. Generally, optical fibers accommodated in an optical cable is provided with a marker band or a plastic ring in the form of a coloring layer or identification code (ID mark), in order to allow identification of the transmission line. When the distance between the optical fibers is reduced, the space available for these protection layers or rings is reduced.
In the production of an optical module that accommodates therein an array of surface-emission laser diodes, it should be noted that the produced optical module would becomes a defective product unless a necessary quality is secured for a predetermined number of laser diode elements in the array. Otherwise, the product loses the value thereof.
This issue is related to the yield of the laser diode production process. In the production of the module product that uses an array arrangement of the laser diode elements, there is an acute demand of establishing the production process in which the modules that function normally are utilized efficiently and the yield of production of the module is improved.
Summarizing above, there is no available long-wavelength surface-emission laser diode operable at the wavelength band of 1.1-1.7 μm and that there is no available optical transmission/reception system that uses such a laser diode.
Also, it is known in the art of surface-emission laser diode to provide a structure in which a current confinement layer (Al2O3) in a part of the p-type semiconductor distributed Bragg reflector close to the active layer by oxidizing an Al(Ga)As selective oxidation layer for the purpose of reducing the threshold current density. It should be noted that the current confinement layer of Al2O3 is a good insulator and the holes constituting the carriers are injected into a limited region of the active layer as a result of the action of the current confinement layer, and it becomes possible to increase the carrier density easily to a threshold carrier density needed for causing laser oscillation. Thereby, it becomes possible to suppress the threshold current to sub milliamperes. Because of the fact that the refractive index of this selective oxidation layer is smaller than the refractive index of the semiconductor layer, the selective oxidation layer functions as an effective optical confinement layer for confining transverse mode, and it becomes possible to obtain a fundamental transverse mode oscillation in the case of reducing the confinement diameter to below about 4 μm in the case of the device is designed for the 0.98 μm band.
In the device in which the confinement diameter is reduced to about 4 μm or less as noted above, on the other hand, there arises a problem of increased electrical resistance because of excessive decrease of the current path area in the current confinement structure. In the device in which the confinement diameter is reduced to the above size or less, for example, it should be noted that the confinement resistance caused as a result of such a current confinement structure constitutes more than the half of the device resistance. As such increase of resistance of the device can become the cause of various problems such as increase of operational voltage, saturation of output power caused by heating, decrease of modulation speed, and the like, it is necessary to reduce the resistance of the confinement structure. This includes not only the reduction of resistance of the current confinement region itself but also the resistance of the peripheral part of the current confinement region.
With regard to the cause of such an increase of resistance as a result of using a current confinement structure, it should be noted that there is a substantial contribution from the high resistance of the p-type semiconductor device used in the p-type semiconductor distributed Bragg reflector. In a semiconductor material, there appears a very high potential barrier at the hetero interface where two semiconductor layers of different bandgaps are contacted, and becomes of this, a p-type semiconductor distributed Bragg reflector generally shows a very high resistance as compared with an n-type semiconductor distributed Bragg reflector.
Conventionally, it is known in the art of surface-emission laser diode of 0.98 μm wavelength to provide a heterospike buffer layer between the two layers having different Al contents and forming a p-type distributed Bragg reflector, for reducing the electric resistance of the distributed Bragg reflector, such that the heterospike buffer layer has a composition intermediate of these two semiconductor layers of different kind. Reference should be made to Photonics Technology Letters, Vol. 2, No. 4, 1990, pp. 234-236, Photonics Technology Letters, Vol. 4, No. 12, 1992, pp. 1325-1327.
Thus, in the art of surface-emission laser diode, decrease of resistance of the device is an important subject matter, and active research and development are being made especially with regard to the reduction of resistance of p-type semiconductor distributed Bragg reflectors. For the desired reduction of the resistance, the use of the hetero barrier buffer layer noted above is extremely effective. Further, it is similarly very effective to increase the doping concentration of the semiconductor layers constituting the semiconductor distributed Bragg reflector, especially the semiconductor layers including the heterospike buffer layer and the layers in the vicinity of the foregoing heterospike buffer layer.
In the case of using a highly doped p-type semiconductor, it is true that the electric characteristics such as device resistance are improved, while there also arise problems such as conspicuous free carrier absorption caused by holes or conspicuous intra-valence band absorption. Thereby, the optical properties of the laser diode are degraded. To improve the electric power transformation efficiency in a surface-emission laser diode, it is particularly important to reduce the absorption of the laser beam by the p-type semiconductor distributed Bragg reflector, while this requirement of reduction of optical absorption loss contradicts with the requirement of reduction of electric resistance.
To eliminate this problem, Japanese Laid-Open Patent Application 2001-332812 proposes a surface-emission laser diode having a semiconductor distributed Bragg reflector in which the doping concentration of the semiconductor distributed Bragg reflector is made relatively low for the region located at the side of the active layer with respect to the region away from the active layer such that the bandgap difference between the two different semiconductor layers of different refractive indices and constituting the semiconductor distributed Bragg reflector is reduced.
In this conventional art, the doping concentration of the semiconductor distributed Bragg reflector located in the vicinity of the active layer is set lower than the doping concentration of other regions, for minimizing the deterioration of the optical output caused by the influence of optical absorption by the semiconductor distributed Bragg reflector. Further, in order to prevent the increase of electric resistance of the semiconductor Bragg reflector caused as a result of reduced doping concentration, the difference of the bandgap is reduced for the semiconductor layers constituting the foregoing less doped region of the semiconductor distributed Bragg reflector such that the potential barrier height formed at the heterointerface is reduced. In the surface-emission laser element having such a construction, the saturation point of the optical output is increased while simultaneously reducing the device resistance.
Thus, in Japanese Laid-Open Patent Application 2001-332812, the doping concentration is reduced in the region located in the vicinity of the active layer in the purpose of reducing the optical absorption, and the bandgap difference between the two semiconductor layers of different kinds and constituting the semiconductor distributed Bragg reflector is reduced for preventing the increase of electric resistance.
However, such a construction, while being able to reduce the resistance to some extent by reducing the bandgap difference for the semiconductor layers constituting the heterointerface, has still suffered from the problem that the electric resistance cannot be reduced sufficiently due to the fact that the reduction of doping concentration inevitably increases the adversary effect of the heterointerface.
Further, the device of Japanese Laid-Open Patent Application 2001-332812 suffers from the problem in that the reduction of bandgap difference leads to decrease of reflectivity of the semiconductor distributed Bragg reflector and penetration of light into the semiconductor distributed Bragg reflector is increased. In order to compensate for this decrease of the reflectivity, it is necessary to increase the number of stacks in the semiconductor distributed Bragg reflector.
Conventionally, it is known in a surface-emission laser diode of the 0.98 μm band, and the like, to provide a hetero barrier buffer layer between the two layers of different Al contents and constituting the distributed Bragg reflector in the form of a compositional graded layer having an Al content intermediate of the foregoing two layers for reducing the electrical resistance of the p-type semiconductor distributed Bragg reflector. Reference should be made to Photonics Technology Letters Vol. 2, No. 4, 1990, pp. 234-236 and Photonics Technology Letters Vol. 4, No. 12, 1992, pp. 1325-1327.
On the other hand, an n-type semiconductor distributed Bragg reflector generally has a very low resistance as compared with a p-type semiconductor distributed Bragg reflector, and no detailed examination have been made so far because it was thought that there would be little influence to the device characteristic (such as device resistance of a surface-emission laser diode).
However, there occurs accumulation or depletion of carriers also in an n-type semiconductor distributed Bragg reflector at the heterointerface formed by two semiconductor layers of different kinds as a result of the influence of band discontinuity between two different semiconductor materials. Because of this, the characteristics of the distributed Bragg reflector differ significantly from those of a bulk semiconductor. Particularly, a depletion layer characterized by decreased carrier density forms electrostatic capacitance component, and become of this, restriction is imposed to electric characteristics and device response characteristics when a device (surface-emission laser diode, and the like) is driven to cause pulse operation or high speed modulation. Furthermore, because of the influence of the heterointerface, there is caused the problem in that non-linearity is caused in the current-voltage characteristic and that the current-voltage characteristic changes in response to the difference of the drive condition of the device.
Thus, it is necessary to conduct detailed examination regarding to the structure and electric characteristics of an n-type semiconductor distributed Bragg reflector in order to obtain a device (surface-emission laser diode, and the like) having excellent characteristics. Conventionally, such detailed examination was not conducted.